Precision peak matching in liquid chromatography-mass spectroscopy

Abstract

A method that identifies common peaks among unidentified peaks in the data from different LC-MS or LC-MS/MS runs is provided. The method employs an algorithm, herein referred to as “Precision Peak Matching (PPM).” The different runs can be from different laboratories, instruments, and biological samples that result in a significant variability in the data. PPM allows estimation and control of precision, defined as the fraction of truly identical peptide pairs among all pairs retrieved, in the matching process. PPM finds the maximal number of peptide pairs at a prescribed precision, thereby allowing quantitative control over the trade off between the number of true pairs missed, and false pairs found. PPM finds common peptides from a database of LC-MS runs of heterogeneous origins, and at the specified precision. PPM fills a much-needed role in proteomics by extracting useful information from disparate LC-MS databases in a statistically rigorous and interpretable manner.

1. A system for matching peaks in liquid chromatography-mass spectroscopy (LC-MS) datasets from multiple runs, said system comprising a memory and a processor device in communication with said memory, wherein said system is configured to perform a method including:
 generating, by employing said processor and said memory, an aligned query list for peaks from a first dataset from a first LC-MS run; 
 generating, by employing said processor and said memory, a target peak list for peaks from a second dataset from a second LC-MS run; 
 generating, by employing said processor and said memory, a mass-to-charge ratio (m/z) tolerance parameter and a retention time (Rt) tolerance parameter that satisfy a specification input criterion for a false matching rate between said aligned query list and said target peak list; 
 determining, by employing said processor and said memory, a true matching rate between said aligned query list and said target peak list employing said m/z tolerance parameter and said Rt tolerance parameter; 
 selecting, by employing said processor and said memory, an optimized m/z tolerance value and an optimized Rt tolerance value by repeating said step of generating said m/z tolerance parameter and Rt tolerance parameter and said step of determining said true matching rate; and 
 generating, by employing said processor and said memory, an optimal list of matches among peaks across said aligned query list and said target peak list employing said optimized m/z tolerance value and said optimized Rt tolerance value as matching parameters. 
 
     
     
       2. The system of  claim 1 , wherein said method further includes:
 fetching, by employing said processor and said memory, said first dataset from an LC-MS apparatus or an LC-MS database; and 
 fetching, by employing said processor and said memory, said second dataset from said LC-MS apparatus, another LC-MS apparatus, said LC-MS database, or another LC-MS database. 
 
     
     
       3. The system of  claim 1 , wherein said step of generating said m/z tolerance parameter and said Rt tolerance parameter includes:
 determining, by employing said processor and said memory, an initial false matching rate between said aligned query list and said target peak list employing an initial m/z tolerance parameter and an initial Rt tolerance parameter; 
 determining, by employing said processor and said memory, whether said initial false matching rate satisfies said specification input criterion; and 
 if said initial false matching rate does not satisfy said specification input criterion, adjusting, by employing said processor and said memory, said initial m/z tolerance parameter and said initial Rt tolerance parameter to generate a revised m/z tolerance parameter and a revised Rt tolerance parameter, respectively, and determining, by employing said processor and said memory, a revised false matching rate between said aligned query list and said target peak list employing said revised m/z tolerance parameter and said revised Rt tolerance parameter. 
 
     
     
       4. The system of  claim 3 , wherein said step of generating said m/z tolerance parameter and said Rt tolerance parameter includes:
 adjusting, by employing said processor and said memory, said revised m/z tolerance parameter and said revised Rt tolerance parameter; 
 determining, by employing said processor and said memory, another revised false matching rate employing a most recent revised m/z tolerance parameter and a most recent revised Rt tolerance parameter; and 
 repeating, by employing said processor and said memory, said step of adjusting said revised m/z tolerance parameter and said revised Rt tolerance parameter and said step of determining said another revised false matching rate until said another revised false matching rate satisfies said specification input criterion. 
 
     
     
       5. The system of  claim 1 , wherein said method further includes:
 comparing, by employing said processor and said memory, a value for said true matching rate from said step of determining said true matching rate with a stored value for said true matching rate from a previously determination of said true matching rate; and 
 if said value for said true matching rate is greater than said stored value, updating, by employing said processor and said memory, said stored value with said value for said true matching rate, storing a value for said m/z tolerance parameter as a stored m/z tolerance parameter, and storing a value for said Rt tolerance parameter as a stored Rt tolerance parameter. 
 
     
     
       6. The system of  claim 5 , wherein said method further includes:
 determining, by employing said processor and said memory, whether an optimization search is complete after comparing said value with said stored value based on a predefined criterion including at least one of a number of iterations at said step of comparing said value with said stored value, a history of said stored value for said true matching rate, a history of said stored m/z tolerance parameter, and a history of said stored Rt tolerance parameter; 
 if said optimization search is complete, generating, by employing said processor and said memory, said optimized m/z tolerance value and said optimized Rt tolerance value, wherein said optimized m/z tolerance value is said stored m/z tolerance parameter and said optimized Rt tolerance value is said stored Rt tolerance parameter. 
 
     
     
       7. The system of  claim 6 , wherein said method further includes conditional steps, wherein said conditional steps include:
 assigning, by employing said processor and said memory, another value for said m/z tolerance parameter and another value for said Rt tolerance parameter; 
 determining, by employing said processor and said memory, another false matching rate employing said m/z tolerance parameter and said Rt tolerance parameter; and 
 determining, by employing said processor and said memory, whether said another false matching rate satisfies said specification input criterion. 
 
     
     
       8. The system of  claim 7 , wherein said conditional steps further include:
 adjusting, by employing said processor and said memory, said m/z tolerance parameter and said revised Rt tolerance parameter until said another false matching rate satisfies said specification input criterion; and 
 repeating, by employing said processor and said memory, said step of determining said true matching rate between said aligned query list and said target peak list employing said m/z tolerance parameter as adjusted and said Rt tolerance parameter as adjusted. 
 
     
     
       9. The system of  claim 1 , wherein said step of determining said true matching rate includes:
 splitting, by employing said processor and said memory, said aligned query list into an annotated query list and a non-annotated query list; 
 splitting, by employing said processor and said memory, said target peak list into an annotated target list and a non-annotated target list; and 
 matching, by employing said processor and said memory, peaks between said annotated query list and said annotated target list. 
 
     
     
       10. The system of  claim 9 , wherein said step of determining, by employing said processor and said memory, said true matching rate further includes generating, by employing said processor and said memory, a true match list based on said matching of peaks between between said annotated query list and said annotated target list, wherein said true matching rate is determined from said true match list. 
     
     
       11. The system of  claim 10 , wherein said true matching rate is a ratio of a total number of matches that does not require a mass shift to achieve a match to a total number of matches in peaks. 
     
     
       12. The system of  claim 1 , wherein said step of generating said m/z tolerance parameter and said Rt) tolerance parameter includes:
 generating, by employing said processor and said memory, a shifted target list by mass-shifting peaks in said target peak list; 
 matching, by employing said processor and said memory, peaks between said aligned query list and said shifter target list; and 
 generating, by employing said processor and said memory, a false match list based on said matching of peaks between said aligned query list and said shifter target list, wherein said false match rate is calculated based on said false match list. 
 
     
     
       13. The system of  claim 12 , wherein said false matching rate is a ratio of a total number of matches that requires a mass shift to achieve a match to a total number of matches in peaks. 
     
     
       14. The system of  claim 12 , wherein said mass-shifting of said peaks in said target peak list shifts said peak by a mass between 3 Da and 200 Da. 
     
     
       15. The system of  claim 1 , wherein said optimal list of matches includes a dataset on peaks, said dataset on peaks including at least a calculated m/z and Rt for each peak.